Can I set the rise period to 0 in JMeter?
I am doing performance testing on a web application.
In the JMeter Tutorial, I learned how to set the Ramp Up period, for example:
If 10 threads are used and the ramp period is 100 seconds, then JMeter will take 100 seconds to start and start all 10 threads. Each stream will start 10 (100/10) seconds after the start of the previous stream.
10 users, 100 second acceleration - start with 1 user, every 10 seconds 1 added user
but I have one doubt, is it possible to set the Ramp Up period to 0? if so, how will the result depend on 0 ramp-up?
source to share
Zero rescission (equivalent to a 1-second spread) is rarely useful, except for large sites that favor release dates, i.e. when a new product is released on a schedule, such as concert tickets or computer games.
You can use it to prove that a site can handle a sudden large influx of requests.
You can probably also find use for stepped build-up, which is available from jmeter-plugins .
As others say, you are trying to simulate normal use, so if normal use involves a sudden heavy load, keep rampup low.
However, keep in mind that with low growth rates and many threads, you may be constrained by local resources, so your results may be a measure of the client's capabilities, not the server's.
Look at distributed load testing to simulate even more workloads than a single client can achieve.
source to share
As stated on the JMeter website, the reason for the ramp-up period is:
The ramp-up should be long enough to avoid too much workload at the beginning of the test, and short enough so that the last threads start before the first ones have finished (unless it does).
As an example, let's say you have a simple test plan that does something like:
- The user is logged in.
- The user adds some content.
- The user is reading some content.
- The user is logged out.
Let's say you configured it to use 1000 threads and have a ramp period of 0, and then when you run your test run it will make 1000 requests at the same time (give or take a bit).
Needless to say, this is not a realistic test. Unless you're experiencing some kind of DDoS, you never expect peaks in usage like this.
Your goal is to mimic a usage profile that is as close as possible to how your system would be used in the real world, and for that to happen, it is likely that you want to distribute usage evenly across all activities in your plan. To do this, you use a rise period. A simple rule of thumb is to set the rise period to be the same as the amount of time you expect from one run of the plan.
source to share